When switching to the 1370 hipass filter, lbw output power would jump and then drift slowing for a few minutes.
A test was done on 10jan19 using:

• lbw linear pol, 1405 Mhz, 750 Mhz mixer.
  
• interim cor 25Mhz bw centered at 1405 Mhz, 1 second dumps. offline compute total power over the 25 Mhz.
  
• lbfb program was used to switch between the lbw rf filters
• lbfb 1 5 : 1370 hipass followed by 1120-1750 filter
• lbfb 6    : 1360 - 1470 filter
• lbfb 3    : 1230-1470 filter
• lbfb 4    : 1550-1820 filter
  

    The lbw rf filter bank has 9 filter selections:

• 7 of these are in an or'ed switch matrix.
• the 1370 hipass filter is switched in out, before the switch matrix
• the 1280 hipass filter is switched in out after the switch matrix.

When using the 1370 hipass filter, you must also select a filter that allows the signal to pass through the switch matrix. In this case i used filter 5 1120-1750 Mhz filter.

    The datataking sequence was:

• adjust the power in the system.
  
• start with the lbfb 1 5 (1370 hp with 1120-1750 filter in).
• data data (about 3 minutes)
• stop datataking
• switch to another filter setting (lbfb 6 or 3 or 4 or 5)
• switch back to lbfb 1 5
• start taking data.
• If the switching worked perfectly, there should be no change in the power levels (except for the sky drifting through the beam).
  

The idea was to see if the signal changed when switching away and the back to the 1370 hipass filter.

The plots show the results of the measurement (.ps) (.pdf):

• The top frame is polA the bottom frame is polB
• The colors: black, red,green,blue are the 4 subbands (through the downstairs mixers) fed to 4 different interim correlator boards.
• These had identical rf input bands.
• The y axis is units of Tsys.
• I normalized each trace to the median value from the 1st 100 sample points of each trace.
• The x axis are the 1 second samples
• The vertical dotted lines are the start of each scan (where the filter was switched away and then back to the 1370 hp).
• At each vertical line you can see jumps in the power level (caused by switching away and then back to the 1370 hp filter.
• After a jump, the power would vary for a few minutes.
  
• the first polA jump was > 8% of Tsys.
• pol B jumps were smaller but occurred more often.
• When leaving Filter 5 in and only switching 1370hp in , out, we still saw a jump in polB power. So the problem is not filter 5.
• The red polA trace was wandering around more than any of the other traces (even when no switches).
  
• This is probably a problem in the downstairs if/lo or interim cor. (more info).
  

Summary

• There is a problem when switching the 1370 hipass filter into the system.
• It looks like the contact is not being made completely. After connection it can drift by up to 5% over a few minutes.
• there is a problem with polA 2nd band downstairs (more info).
• I don't know how long this problem has been present.
  

processing: x101/190110/lbwchk.pro